Searching the sky with CONFIGR-STARS

CONFIGR-STARS, a new methodology based on a model of the human visual system, is developed for registration of star images. The algorithm first applies CONFIGR, a neural model that connects sparse and noisy image components. CONFIGR produces a web of connections between stars in a reference starmap or in a test patch of unknown location. CONFIGR-STARS splits the resulting, typically highly connected, web into clusters, or “constellations”. Cluster geometry is encoded as a signature vector that records edge lengths and angles relative to the cluster’s baseline edge. The location of a test patch cluster is identified by comparing its signature to signatures in the codebook of a reference starmap, where cluster locations are known. Simulations demonstrate robust performance in spite of image perturbations and omissions, and across starmaps from different sources and seasons. Further studies would test CONFIGR-STARS and algorithm variations applied to very large starmaps and to other technologies that may employ geometric signatures. Open-source code, data, and demos are available from http://techlab.bu.edu/STARS/.     

Effects of selective attention on perceptual filling-in

After few seconds, a figure steadily presented in peripheral vision becomes perceptually filled-in by its background, as if it "disappeared". We report that directing attention to the color, shape, or location of a figure increased the probability of perceiving filling-in compared to unattended figures, without modifying the time required for filling-in. This effect could be augmented by boosting attention. Furthermore, the frequency distribution of filling-in response times for attended figures could be predicted by multiplying the frequencies of response times for unattended figures with a constant. We propose that, after failure of figure-ground segregation, the neural interpolation processes that produce perceptual filling-in are enhanced in attended figure regions. As filling-in processes are involved in surface perception, the present study demonstrates that even very early visual processes are subject to modulation by cognitive factors.     

Invariant template matching in systems with spatiotemporal coding: A matter of instability

We consider the design principles of algorithms that match templates to images subject to spatiotemporal encoding. Both templates and images are encoded as temporal sequences of samplings from spatial patterns. Matching is required to be tolerant to various combinations of image perturbations. These include ones that can be modeled as parameterized uncertainties such as image blur, luminance, and, as special cases, invariant transformation groups such as translation and rotations, as well as unmodeled uncertainties (noise). For a system to deal with such perturbations in an efficient way, they are to be handled through a minimal number of channels and by simple adaptation mechanisms. These normative requirements can be met within the mathematical framework of weakly attracting sets. We discuss explicit implementation of this principle in neural systems and show that it naturally explains a range of phenomena in biological vision, such as mental rotation, visual search, and the presence of multiple time scales in adaptation. We illustrate our results with an application to a realistic pattern recognition problem.     

Recovering real-world images from single-scale boundaries with a novel filling-in architecture

Filling-in models were successful in predicting psychophysical data for brightness perception. Nevertheless, their suitability for real-world image processing has never been examined. A unified architecture for both predicting psychophysical data and real-world image processing would constitute a powerful theory for early visual information processing. As a first contribution of the present paper, we identified three principal problems with current filling-in architectures, which hamper the goal of having such a unified architecture. To overcome these problems we propose an advance to filling-in theory, called BEATS filling-in, which is based on a novel nonlinear diffusion operator. BEATS filling-in furthermore introduces novel boundary structures. We compare, by means of simulation studies with real-world images, the performance of BEATS filling-in with the recently proposed confidence-based filling-in. As a second contribution we propose a novel mechanism for encoding luminance information in contrast responses (‘multiplex contrasts’), which is based on recent neurophysiological findings. Again, by simulations, we show that ‘multiplex contrasts’ at a single, high-resolution filter scale are sufficient for recovering absolute luminance levels. Hence, ‘multiplex contrasts’ represent a novel theory addressing how the brain encodes and decodes luminance information.     

Principles of visual motion detection

Motion information is required for the solution of many complex tasks of the visual system such as depth perception by motion parallax and figure/ground discrimination by relative motion. However, motion information is not explicitly encoded at the level of the retinal input. Instead, it has to be computed from the time-dependent brightness patterns of the retinal image as sensed by the two-dimensional array of photoreceptors. Different models have been proposed which describe the neural computations underlying motion detection in various ways. To what extent do biological motion detectors approximate any of these models? As will be argued here, there is increasing evidence from the different disciplines studying biological motion vision, that, throughout the animal kingdom ranging from invertebrates to vertebrates including man, the mechanisms underlying motion detection can be attributed to only a few, essentially equivalent computational principles. Motion detection may, therefore, be one of the first examples in computational neurosciences where common principles can be found not only at the cellular level (e.g., dendritic integration, spike propagation, synaptic transmission) but also at the level of computations performed by small neural networks.     

Fast Image Recognition Based on Independent Component Analysis and Extreme Learning Machine

Nowadays, image recognition has become a highly active research topic in cognitive computation community, due to its many potential applications. Generally, the image recognition task involves two subtasks: image representation and image classification. Most feature extraction approaches for image representation developed so far regard independent component analysis (ICA) as one of the essential means. However, ICA has been hampered by its extremely expensive computational cost in real-time implementation. To address this problem, a fast cognitive computational scheme for image recognition is presented in this paper, which combines ICA and the extreme learning machine (ELM) algorithm. It tries to solve the image recognition problem at a much faster speed by using ELM not only in image classification but also in feature extraction for image representation. As an example, our proposed approach is applied to the face image recognition with detailed analysis. Firstly, common feature hypothesis is introduced to extract the common visual features from universal images by the traditional ICA model in the offline recognition process, and then ELM is used to simulate ICA for the purpose of facial feature extraction in the online recognition process. Lastly, the resulting independent feature representation of the face images extracted by ELM rather than ICA will be fed into the ELM classifier, which is composed of numerous single hidden layer feed-forward networks. Experimental results on Yale face database and MNIST digit database have shown the good performance of our proposed approach, which could be comparable to the state-of-the-art techniques at a much faster speed.     

Electrophysiological correlates of similarity-based interference during detection of visual forms

Illusory figure completion demonstrates the ability of the visual system to integrate information across gaps. Mechanisms that underlie figural emergence support the interpolation of contours and the filling-in of form information [Grossberg, S., & Mingolla, E. Neural dynamics of form perception: Boundary completion, illusory figures and neon colour spreading. Psychological Review, 92, 173-211, 1985]. Although both processes contribute to figure formation, visual search for an illusory target configuration has been shown to be susceptible to interfering form, but not contour, information [Conci, M., Müller, H. J., & Elliott, M. A. The contrasting impact of global and local object attributes on Kanizsa figure detection. Submitted]. Here, the physiological basis of form interference was investigated by recording event-related potentials elicited from contour- and surface-based distracter interactions with detection of a target Kanizsa figure. The results replicated the finding of form interference and revealed selection of the target and successful suppression of the irrelevant distracter to be reflected by amplitude differences in the N2pc component (240-340 msec). In conclusion, the observed component variations reflect processes of target selection on the basis of integrated form information resulting from figural completion processes.     

Fast Face Recognition Via Sparse Coding and Extreme Learning Machine

Most face recognition approaches developed so far regard the sparse coding as one of the essential means, while the sparse coding models have been hampered by the extremely expensive computational cost in the implementation. In this paper, a novel scheme for the fast face recognition is presented via extreme learning machine (ELM) and sparse coding. The common feature hypothesis is first introduced to extract the basis function from the local universal images, and then the single hidden layer feedforward network (SLFN) is established to simulate the sparse coding process for the face images by ELM algorithm. Some developments have been done to maintain the efficient inherent information embedding in the ELM learning. The resulting local sparse coding coefficient will then be grouped into the global representation and further fed into the ELM ensemble which is composed of a number of SLFNs for face recognition. The simulation results have shown the good performance in the proposed approach that could be comparable to the state-of-the-art techniques at a much higher speed.     

A No-Reference Image Quality Measure for Blurred and Compressed Images Using Sparsity Features

Images can be distorted in the real world via many sources like faulty sensors, artifacts generated by compression algorithms, defocus, faulty lens, and poor lighting conditions. Our biological vision system can identify the quality of image by looking at the images, but developing an algorithm to assess the quality of an image is a very challenging task as an image can be corrupted by different types of distortions and statistical properties of different types of distortions are dissimilar. The main objective of this article is to propose an image quality assessment technique for images corrupted by blurring and compression-based artifacts. Machine learning-based approaches have been used in recent times to perform this task. Images can be analyzed in different transform domains like discrete cosine transform domain, wavelet domains, curvelet domains, and singular value decomposition. These domains generate sparse matrices. In this paper, we propose no-reference image quality assessment algorithms for images corrupted by blur and different compression algorithms using sparsity-based features computed from different domains and all features pooled by support vector regression. The proposed model has been tested on three standard image quality assessment datasets LIVE, CSIQ, and TID2013, and correlation with subjected human opinion scores has been presented along with comparative study with state-of-the-art quality measures. Experiments run on standard image quality databases show that the results obtained are outperforming the existing results.     

A neural network architecture for figure-ground separation of connected scenic figures

A neural network model, called an FBF network, is proposed for automatic parallel separation of multiple image figures from each other and their backgrounds in noisy gray-scale or multicolored images. The figures can then be processed in parallel by an array of self-organizing Adaptive Resonance Theory (ART) neural networks for automatic target recognition. An FBF network can automatically separate the disconnected but interleaved spirals that Minsky and Papert introduced in their book Perceptrons. The network's design also clarifies why humans cannot rapidly separate interleaved spirals, yet can rapidly detect conjunctions of disparity and color, or of disparity and motion, that distinguish target figures from surrounding distractors. Figure-ground separation is accomplished by iterating operations of a Feature Contour System (FCS) and a Boundary Contour System (BCS) in the order FCS-BCS-FCS, hence the term FBF. The FCS operations include the use of nonlinear shunting networks to compensate for variable illumination and nonlinear diffusion networks to control filling-in. A key new feature of an FBF networks is the use of filling-in for figure-ground separation. The BCS operations include oriented filters joined to competitive and cooperative interactions designed to detect, regularize, and complete boundaries in up to 50% noise, while suppressing the noise. A modified CORT-X filter is described. which uses both on-cells and off-cells to generate a boundary segmentation from a noisy image.     

3D Object Recognition Based on Some Aspects of the Infant Vision System and Associative Memory

A view-based method for 3D object recognition based on some biological aspects of infant vision is proposed in this paper. The biological hypotheses of this method are based on the role of the response to low frequencies at early stages as well as some conjectures concerning how an infant detects subtle features (stimulating points) from an object. In order to recognize an object from different images of it (at different orientations from 0° to 360°), we make use of a dynamic associative memory (DAM). As the infant vision responds to low frequencies of the signal, a low-filter is first used to remove high frequency components from the image. Then, we detect subtle features in the image by means of a random feature selection detector. At last, the DAM is fed with this information for training and recognition. To test the accuracy of the proposed model, we use the Columbia Object Image Library (COIL 100) database.     

The effectiveness of FID-CT (functional image of dynamic CT) for the evaluation of early cerebral ischemia

A computer program was newly developed to display the ischemic area from the serial dynamic CT scan, namely the functional image of the dynamic CT (FID-CT). The principles of FID-CT are as follows; Seven rapid-sequence dynamic CT scans were taken following a peripheral bolus intravenous injection of 40 ml of iopamidol. As the data from each scan can be separated into three consecutive segments, we obtained 21 images during 44 seconds. Time density curves of each pixel were calculated employing Gamma variate fitting method. Eight functional parameters obtained from this curve were displayed with gray-scale pixel by pixel. These eight images were functional images obtained from dynamic CT. Only three or four minutes were required to complete all the calculations. Twenty-two patients were examined with both FID-CT and 123I-SPECT. In each case, the lesion detected by FID-CT was remarkably consistent with that shown by SPECT. Two representative cases were presented. The authors believe that FID-CT is a very useful diagnostic method in the acute stage of cerebral ischemia because the method can be quickly and easily performed and it discloses the ischemic area with fair certainty.     

Color constancy and the functional significance of McCollough effects.

A central problem in visual perception concerns how humans perceive stable and uniform object colors despite variable lighting conditions (i.e. color constancy). One solution is to 'discount' variations in lighting across object surfaces by encoding color contrasts, and utilize this information to 'fill in' properties of the entire object surface. Implicit in this solution is the caveat that the color contrasts defining object boundaries must be distinguished from the spurious color fringes that occur naturally along luminance-defined edges in the retinal image (i.e. optical chromatic aberration). In the present paper, we propose that the neural machinery underlying color constancy is complemented by an 'error-correction' procedure which compensates for chromatic aberration, and suggest that error-correction may be linked functionally to the experimentally induced illusory colored aftereffects known as McCollough effects (MEs). To test these proposals, we develop a neural network model which incorporates many of the receptive-field (RF) profiles of neurons in primate color vision. The model is composed of two parallel processing streams which encode complementary sets of stimulus features: one stream encodes color contrasts to facilitate filling-in and color constancy; the other stream selectively encodes (spurious) color fringes at luminance boundaries, and learns to inhibit the filling-in of these colors within the first stream. Computer simulations of the model illustrate how complementary color-spatial interactions between error-correction and filling-in operations (a) facilitate color constancy, (b) reveal functional links between color constancy and the ME, and (c) reconcile previously reported anomalies in the local (edge) and global (spreading) properties of the ME. We discuss the broader implications of these findings by considering the complementary functional roles performed by RFs mediating color-spatial interactions in the primate visual system.     

Neural correlates of hemianopic completion across the vertical meridian

Hemianopic completion refers to the perceptual completion of figures located across the vertical meridian in the context of hemianopia, such that one half of the figure falls within the blind hemifield. It can occur whether the figure is itself complete (veridical completion) or incomplete (paracompletion). Psychophysical evidence suggests that this phenomenon may be a constructive one, and may share features with completion phenomena in normal vision. The neural structures mediating hemianopic completion are unknown. Here we studied the neural activity evoked by hemianopic completion using event-related fMRI in an individual (POV) with a large right visual field homonymous hemianopic scotoma due to left occipital damage. Either a large achromatic circular contour straddling the vertical meridian or a semicircular contour within the left hemifield just crossing the vertical meridian was presented to POV on each trial. POV indicated by button press whether he perceived a semicircular contour, a patchy circular contour or a complete circular contour. On trials where he reported perceiving a complete circular contour despite being presented with a semicircular contour (paracompletion), activity was increased in a region of ipsilateral extrastriate cortex (contralateral to the lesion, ipsilateral to the illusory edge of the circle). These results are discussed in the context of illusory contour completion in healthy subjects and more generally in the recovery of function after brain damage.     

Learning invariant object recognition from temporal correlation in a hierarchical network

Invariant object recognition, which means the recognition of object categories independent of conditions like viewing angle, scale and illumination, is a task of great interest that humans can fulfill much better than artificial systems. During the last years several basic principles were derived from neurophysiological observations and careful consideration: (1) Developing invariance to possible transformations of the object by learning temporal sequences of visual features that occur during the respective alterations. (2) Learning in a hierarchical structure, so basic level (visual) knowledge can be reused for different kinds of objects. (3) Using feedback to compare predicted input with the current one for choosing an interpretation in the case of ambiguous signals. In this paper we propose a network which implements all of these concepts in a computationally efficient manner which gives very good results on standard object datasets. By dynamically switching off weakly active neurons and pruning weights computation is sped up and thus handling of large databases with several thousands of images and a number of categories in a similar order becomes possible. The involved parameters allow flexible adaptation to the information content of training data and allow tuning to different databases relatively easily. Precondition for successful learning is that training images are presented in an order assuring that images of the same object under similar viewing conditions follow each other. Through an implementation with sparse data structures the system has moderate memory demands and still yields very good recognition rates.     

Analysis of the process of visual pattern recognition by the neocognitron

A neural network model of visual pattern recognition called the “neocognitron,” was earlier proposed by the author. It is capable of deformation-invariant visual pattern recognition. After learning, it can recognize input patterns without being affected by deformation, changes in size, or shifts in position. This paper offers a mathematical analysis of the process of visual pattern recognition by the neocognitron. The neocognitron is a hierarchical multilayered network. Its initial stage is an input layer, and each succeeding stage has a layer of “S-cells” followed by a layer of “C-cells.” Thus, in the whole network, layers of S-cells and C-cells are arranged alternately. The process of feature extraction by an S-cell is analyzed mathematically in this paper, and the role of the C-cells in deformation-invariant pattern recognition is discussed.     

Anatomical standardization of small animal brain FDG-PET images using synthetic functional template: experimental comparison with anatomical template

Anatomical standardization (also called spatial normalization) of positron emission tomography (PET) small animal brain images is required to make statistical comparisons across individuals. Frequently, PET images are co-registered to an individual MR or CT image of the same subject in order to transform the functional images to an anatomical space. In the present work, we evaluate the normalization of synthetic PET (synPET) images to a synthetic PET template. To provide absolute error in terms of pixel misregistration, we created a synthetic PET image from the individual MR image through segmentation of the brain into gray and white matter which produced functional and anatomical images in the same space. When comparing spatial normalization of synPET images to a synPET template with the gold standard (MR images to an MR template), a mean translation error of 0.24mm (±0.20) and a maximal mean rotational error of 0.85° (±0.91) were found. Significant decrease in misregistration error was measured when achieving spatial normalization of functional images to a functional template instead of an anatomical template. This accuracy strengthens the use of standardization methods where individual PET images are registered to a customized PET template in order to statistically assess physiological changes in rat brains.     

图像数据挖掘的模型和技术

目的：介绍图像数据挖掘的模型及核心技术。 方法：原始图像不能直接用于图像数据挖掘,必须进行预处理以生成用于高层次挖掘的图像特征库。一个图像挖掘系统应该包括图像的存储、预处理、检索、挖掘和展示等功能。它主要涉及图像数据挖掘模型和图像数据挖掘技术。 结果：MultiMediaMiner是以DBMiner系统和C-BIRD系统为基础发展起来的图像数据挖掘系统,它是典型的功能驱动模型。在信息驱动模型中,象素层和对象层主要进行图像处理、对象识别和特征提取,而语义概念层和模式知识层主要进行图像数据挖掘和知识集成,该模型不仅只在图像的高层次进行挖掘,而且还可以扩展此模型以使挖掘能够在每个层次以及不同层次间进行。基于图像的数据挖掘核心技术涉及：图像处理技术,如去噪、对比度增强、图像分割等技术;特征提取和优化技术;分类、规则提取、预测和聚类等。 结论：理论上图像数据挖掘是数据挖掘的一个分支,但是由于挖掘对象的复杂性,所以图像数据挖掘不是传统的数据挖掘理论与技术在图像数据上的简单应用和延伸,而是一个具有自己独特研究内容、理论与技术框架的新的研究领域。     

Computational principles of learning in the neocortex and hippocampus

We present an overview of our computational approach towards understanding the different contributions of the neocortex and hippocampus in learning and memory. The approach is based on a set of principles derived from converging biological, psychological, and computational constraints. The most central principles are that the neocortex employs a slow learning rate and overlapping distributed representations to extract the general statistical structure of the environment, while the hippocampus learns rapidly, using separated representations to encode the details of specific events while suffering minimal interference. Additional principles concern the nature of learning (error-driven and Hebbian), and recall of information via pattern completion. We summarize the results of applying these principles to a wide range of phenomena in conditioning, habituation, contextual learning, recognition memory, recall, and retrograde amnesia, and we point to directions of current development.     

Spatial contrast sensitivity in facial recognition

Recognition and learning of complex images may depend on spatial processing characteristics of the visual system. Prosopagnosia, an impairment of visual learning and recognition of faces, might result from impaired perception in spatial channels carrying crucial information. We studied spatial contrast sensitivity (SCS) in two subjects with stable facial recognition defects. One had relative SCS reduction for high-frequency gratings but could process high frequencies in room light. The other had normal SCS. Both had intact spatial processing relative to image size. The results suggest that impairments in visual spatial channels are not necessary for the development of prosopagnosia.